import os
from transformers import pipeline
from gtts import gTTS
from simple_salesforce import Salesforce
import soundfile as sf

# Step 1: Hugging Face Speech-to-Text Pipeline Setup
speech_to_text = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-large-960h")

# Step 2: Convert Speech to Text
def convert_speech_to_text(audio_file):
    """ Convert audio file to text """
    try:
        with open(audio_file, "rb") as audio:
            transcription = speech_to_text(audio.read())
            return transcription['text']
    except Exception as e:
        print(f"Error in converting speech to text: {e}")
        return None

# Step 3: Analyze the text for health-related indicators (e.g., respiratory issues)
health_assessment = pipeline("zero-shot-classification", model="facebook/bart-large-mnli")

# Define possible health conditions
health_conditions = ["respiratory issues", "mental health conditions", "fever", "asthma", "coughing"]

def analyze_health_condition(text):
    """ Analyze the health condition based on transcribed text """
    try:
        result = health_assessment(text, candidate_labels=health_conditions)
        return result
    except Exception as e:
        print(f"Error in analyzing health condition: {e}")
        return None

# Step 4: Provide Feedback to the User Based on Health Assessment
def provide_feedback(health_assessment_result):
    """ Provide feedback based on health assessment """
    if health_assessment_result is None:
        return "Error in analyzing health, please try again later."

    try:
        if 'respiratory issues' in health_assessment_result['labels']:
            return "Possible respiratory issue detected, consult a doctor."
        elif 'mental health conditions' in health_assessment_result['labels']:
            return "Possible mental health concern detected, seek professional help."
        else:
            return "No significant health concerns detected. Keep monitoring your health."
    except Exception as e:
        print(f"Error in providing feedback: {e}")
        return "An error occurred while processing your health assessment."

# Step 5: Convert Text Feedback to Speech (Text-to-Speech)
def text_to_speech(text):
    """ Convert text feedback to speech """
    try:
        tts = gTTS(text, lang='en')
        tts.save("response.mp3")
        os.system("start response.mp3")  # Play the audio file (Windows-specific command)
    except Exception as e:
        print(f"Error in text to speech: {e}")

# Step 6: Integration with Salesforce for Storing User Data
def store_user_data_to_salesforce(user_first_name, user_last_name, user_email, feedback):
    """ Store user data to Salesforce """
    try:
        # Salesforce login credentials
        sf = Salesforce(username='your_username', password='your_password', security_token='your_token')

        # Create a new record for the user interaction in Salesforce
        sf.Contact.create({
            'FirstName': user_first_name,
            'LastName': user_last_name,
            'Email': user_email,
            'VoiceAnalysisResult': feedback,
        })
        print(f"Data stored successfully for {user_first_name} {user_last_name}.")
    except Exception as e:
        print(f"Error in storing data to Salesforce: {e}")

# Step 7: Main Function to Process User's Voice Input
def analyze_voice_health(audio_file, user_first_name, user_last_name, user_email):
    """ Main function to analyze voice and provide feedback """
    # Step 1: Convert speech to text
    text = convert_speech_to_text(audio_file)
    if text is None:
        return "Error in transcribing the audio."

    print(f"User's Speech Transcription: {text}")

    # Step 2: Analyze the transcribed text for health conditions
    health_feedback = analyze_health_condition(text)
    if health_feedback is None:
        return "Error in analyzing health conditions."

    print(f"Health assessment: {health_feedback}")

    # Step 3: Provide feedback based on the health analysis
    feedback = provide_feedback(health_feedback)
    print(f"Feedback: {feedback}")

    # Step 4: Convert the feedback to speech for accessibility
    text_to_speech(feedback)

    # Step 5: Store the user interaction data in Salesforce (optional)
    store_user_data_to_salesforce(user_first_name, user_last_name, user_email, feedback)
    
    return feedback


# Example Usage:
# Assuming you have a user's voice recorded in "user_voice.wav"
audio_file = "user_voice.wav"
user_first_name = "John"
user_last_name = "Doe"
user_email = "johndoe@example.com"

# Analyze the user's voice and provide feedback
feedback = analyze_voice_health(audio_file, user_first_name, user_last_name, user_email)
print(f"Final Feedback: {feedback}")